Search

The increasing use of underwater vehicles facilitates deep-sea exploration at a wide range of depths and spatial scales. In this paper, we make an initial attempt to develop online computing strategies to identify seafloor categories and predict biogeographic patterns with a deep learning-based architecture, DenseNet, integrated with joint morphological cues, with the expectation of potentially developing its embedded smart capacities. We utilized high-resolution multibeam bathymetric measurements derived from MBES and denoted a collection of joint morphological cues to help with semantic mapping and localization. We systematically strengthened dominant feature propagation and promoted feature reuse via DenseNet by applying the channel attention module and spatial pyramid pooling. From our experiment results, the seafloor classification accuracy reached up to 89.87%, 82.01%, and 73.52% on average in terms of PA, MPA, and MIoU metrics, achieving comparable performances with the state-of-the-art deep learning frameworks. We made a preliminary study on potential biogeographic distribution statistics, which allowed us to delicately distinguish the functionality of probable submarine benthic habitats. This study demonstrates the premise of using underwater vehicles through unbiased means or pre-programmed path planning to quantify and estimate seafloor categories and the exhibited fine-scale biogeographic patterns.

This paper proposes a prediction strategy for the hydrodynamic performance of bionic fish. The major challenges are meshing and building prediction model. The NACA0012 airfoil is used to replace the fish driven by the body and/or caudal fin (BCF), and a two-dimensional swimming geometric model is constructed. The geometric model is divided into hybrid meshes using the overset mesh method. The classical traveling wave model is studied using Matlab, and an improved self-propelled motion model is established. The hydrodynamic performance of the self-propelled model is numerically simulated based on computational fluid dynamics (CFD). In this strategy, the geometric model and the self-propelled motion model are integrated with user defined functions (UDF). The influences of the parameters such as inflow velocity, frequency, wavelength, and head fluctuation amplitude on the motion performance are studied. The results show that when inflow velocity is uniform, the self-propelled motion will eventually reach a quasi-steady state. According to the numerical simulation results, a hydrodynamic performance prediction model is established based on multilayer perceptron (MLP). The model is used to predict the performance of the optimized traveling wave parameters, and the error is within 3[Formula: see text]. The accuracy and generalization ability of the MLP prediction model is verified.

Lidar-based 3D object detection and classification is a critical task for autonomous driving. However, inferencing from exceedingly sparse 3D data in real-time is a formidable challenge. Complex-YOLO solves the problem of point cloud disorder and sparsity by projecting it onto the bird’s-eye view and realizes real-time 3D object detection based on LiDAR. However, Complex-YOLO has no object height detection, a shallow network depth, and poor small-size object detection accuracy. To address these issues, this paper has made the following improvements: (1) adds a multi-scale feature fusion network to improve the algorithm’s capability to detect small-size objects; (2) uses a more advanced RepVGG as the backbone network to improve network depth and overall detection performance; and (3) adds an effective height detector to the network to improve the height detection. Through experiments, we found that our algorithm’s accuracy achieved good performance on the KITTI dataset, while the detection speed and memory usage were very superior, 48FPS on RTX3070Ti and 20FPS on GTX1060, with a memory usage of 841Mib.

Based on data-driven and mixed models, this study proposes a fault detection method for autonomous underwater vehicle (AUV) rudder systems. The proposed method can effectively detect faults in the absence of angle feedback from the rudder. Considering the parameter uncertainty of the AUV motion model resulting from the dynamics analysis method, we present a parameter identification method based on the recurrent neural network (RNN). Prior to identification, singular value decomposition (SVD) was chosen to denoise the original sensor data as the data pretreatment step. The proposed method provides more accurate predictions than recursive least squares (RLSs) and a single RNN. In order to reduce the influence of sensor parameter errors and prediction model errors, the adaptive threshold is mentioned as a method for analyzing prediction errors. In the meantime, the results of the threshold analysis were combined with the qualitative force analysis to determine the rudder system’s fault diagnosis and location. Experiments conducted at sea demonstrate the feasibility and effectiveness of the proposed method.

Ego-motion estimation, as one of the core technologies of unmanned systems, is widely used in autonomous robot navigation, unmanned driving, augmented reality and other fields. With the development of computer vision, there has been considerable interest in ego-motion estimation with visual navigation. One of the core technologies in Visual navigation is using the matching feature points between consecutive image frames to estimate pose. Since the feature-based method performed under the assumption of a static environment, it susceptive to the dynamic targets. Visual navigation in the dynamic environment has become an important research issue. This paper proposed a practical and robust features selection algorithm of visual navigation which avoids using the feature points on dynamic objects. Firstly, according to the instance segmentation of deep neural network, the objects are classified into potential dynamic and static categories. Subsequently, the matching features on the potential moving objects are used to update vehicle state respectively, meanwhile, the relevant reprojection error of other feature points on the background could be calculated. Eventually, the result of whether the target is moving or not will be judged by the reprojection error, and the features on dynamic targets are removed. To illustrate the effectiveness of the features selection method in the dynamic environment, the proposed algorithm is merged into an MSCKF based on tri-focal tensor geometry, and it has been evaluated in a public dataset. Experimental results demonstrated the effectiveness of the proposed method.

Nowadays, industrial processes are fully equipped by redundant hardware sensors which can be interfered by random noises. Hence, it is of high importance to develop soft sensing solutions for key variables prediction and process monitoring. Various methods have been carried out to cope with different data characteristics among which auto-correlation and non-stable features have been considered as two challenging tasks. In this paper, a novel weighted autoregressive dynamic latent variable (WARDLV) model is constructed where a weighted log-likelihood function is created combined with high-order dynamic information extraction techniques. Different weights are designed for state transition and emission in latent space and original space, respectively. Compared with previous researches, global weights are completely redesigned and consisted of different window confidences and local weights. The efficiency of the proposed method is further demonstrated with a real industrial application.

Purpose: Wavelet denoising is one of the denoising methods commonly used for ECG signals. However, due to the frequency overlap between the EMG and ECG, the feeble characteristics of ECG signals exists the risk of being weakened in the process of filtering noise. This paper presents a method of modified wavelet design and applies it to the denoising of ECG signals. Materials and methods: The optimized filter coefficients are obtained by approximating the amplitude-frequency response of the ideal filter, and the wavelet is constructed with the optimized filter coefficients. The algorithm is tested by clinical ECG data. Results: The results show that the proposed denoising method can remove the high-frequency noise effectively and enhance the characteristic information of P waves and T waves, and retain the characteristic information of the atrial fibrillation signals simultaneously. Compared with db4 and sym4 wavelets, the proposed wavelet can improve the signal to noise ratio and reduce the mean square error effectively at the same time. Conclusion: The modified wavelet design method proposed in this paper can effectively remove high-frequency noise while retaining and enhancing weak features. It provides a theoretical guidance for the de-noising of ECG signals in mobile medicine and also provides a way for other types of weak feature signal denoising.

Aiming to improve the performance of motion for autonomous underwater vehicle (AUV), a fractional-order PID strategy is proposed. It is a more generalized form for the conventional integer-order PID controller, keeping its simplicity and utilizing the generalized derivative and integral control actions. The fractional-order PID controller has been successfully applied to heading control, diving control and path-following system of AUV on sea trial. In addition, the fractional-order closed-loop system has proven to be stable. By comparing simulations and experiments, the satisfactory performance, such as overshoot, settling time and steady-state error, has been achieved. The cloud-model-based quantum genetic algorithm (CQGA) is employed to tune coefficients of fractional-order PID controller. The quantum bits and quantum superposition states avoid the pressure of selection and maintain the diversity of population in chromosome coding. Due to the randomness and stability tendency of cloud droplets, the cloud crossover operator and the cloud mutation operator can effectively overcome the shortcomings of premature and slow searching speed. Numerical simulations show that the CQGA is more efficient to find the optimal coefficients of fractional-order PID controller than GA.

Underwater fishing nets represent a danger faced by autonomous underwater vehicles (AUVs). To avoid irreparable damage to the AUV caused by fishing nets, the AUV needs to be able to identify and locate them autonomously and avoid them in advance. Whether the AUV can avoid fishing nets successfully depends on the accuracy and efficiency of detection. In this paper, we propose an object detection multiple receptive field network (MRF-Net), which is used to recognize and locate fishing nets using forward-looking sonar (FLS) images. The proposed architecture is a center-point-based detector, which uses a novel encoder-decoder structure to extract features and predict the center points and bounding box size. In addition, to reduce the interference of reverberation and speckle noises in the FLS image, we used a series of preprocessing operations to reduce the noises. We trained and tested the network with data collected in the sea using a Gemini 720i multi-beam forward-looking sonar and compared it with state-of-the-art networks for object detection. In order to further prove that our detector can be applied to the actual detection task, we also carried out the experiment of detecting and avoiding fishing nets in real-time in the sea with the embedded single board computer (SBC) module and the NVIDIA Jetson AGX Xavier embedded system of the AUV platform in our lab. The experimental results show that in terms of computational complexity, inference time, and prediction accuracy, MRF-Net is better than state-of-the-art networks. In addition, our fishing net avoidance experiment results indicate that the detection results of MRF-Net can support the accurate operation of the later obstacle avoidance algorithm.

In this paper, a novel iterative sparse extended information filter (ISEIF) was proposed to solve the simultaneous localization and mapping problem (SLAM), which is very crucial for autonomous vehicles. The proposed algorithm solves the measurement update equations with iterative methods adaptively to reduce linearization errors. With the scalability advantage being kept, the consistency and accuracy of SEIF is improved. Simulations and practical experiments were carried out with both a land car benchmark and an autonomous underwater vehicle. Comparisons between iterative SEIF (ISEIF), standard EKF and SEIF are presented. All of the results convincingly show that ISEIF yields more consistent and accurate estimates compared to SEIF and preserves the scalability advantage over EKF, as well.

This paper presents a novel and practical convolutional neural network architecture to implement semantic segmentation for side scan sonar (SSS) image. As a widely used sensor for marine survey, SSS provides higher-resolution images of the seafloor and underwater target. However, for a large number of background pixels in SSS image, the imbalance classification remains an issue. What is more, the SSS images contain undesirable speckle noise and intensity inhomogeneity. We define and detail a network and training strategy that tackle these three important issues for SSS images segmentation. Our proposed method performs image-to-image prediction by leveraging fully convolutional neural networks and deeply-supervised nets. The architecture consists of an encoder network to capture context, a corresponding decoder network to restore full input-size resolution feature maps from low-resolution ones for pixel-wise classification and a single stream deep neural network with multiple side-outputs to optimize edge segmentation. We performed prediction time of our network on our dataset, implemented on a NVIDIA Jetson AGX Xavier, and compared it to other similar semantic segmentation networks. The experimental results show that the presented method for SSS image segmentation brings obvious advantages, and is applicable for real-time processing tasks.

Navigation technology is one of the most important challenges in the applications of autonomous underwater vehicles (AUVs) which navigate in the complex undersea environment. The ability of localizing a robot and accurately mapping its surroundings simultaneously, namely the simultaneous localization and mapping (SLAM) problem, is a key prerequisite of truly autonomous robots. In this paper, a modified-FastSLAM algorithm is proposed and used in the navigation for our C-Ranger research platform, an open-frame AUV. A mechanical scanning imaging sonar is chosen as the active sensor for the AUV. The modified-FastSLAM implements the update relying on the on-board sensors of C-Ranger. On the other hand, the algorithm employs the data association which combines the single particle maximum likelihood method with modified negative evidence method, and uses the rank-based resampling to overcome the particle depletion problem. In order to verify the feasibility of the proposed methods, both simulation experiments and sea trials for C-Ranger are conducted. The experimental results show the modified-FastSLAM employed for the navigation of the C-Ranger AUV is much more effective and accurate compared with the traditional methods.

This paper addresses an autonomous navigation method for the autonomous underwater vehicle (AUV) C-Ranger applying information-filter-based simultaneous localization and mapping (SLAM), and its sea trial experiments in Tuandao Bay (Shangdong Province, P.R. China). Weak links in the information matrix in an extended information filter (EIF) can be pruned to achieve an efficient approach-sparse EIF algorithm (SEIF-SLAM). All the basic update formulae can be implemented in constant time irrespective of the size of the map; hence the computational complexity is significantly reduced. The mechanical scanning imaging sonar is chosen as the active sensing device for the underwater vehicle, and a compensation method based on feedback of the AUV pose is presented to overcome distortion of the acoustic images due to the vehicle motion. In order to verify the feasibility of the navigation methods proposed for the C-Ranger, a sea trial was conducted in Tuandao Bay. Experimental results and analysis show that the proposed navigation approach based on SEIF-SLAM improves the accuracy of the navigation compared with conventional method; moreover the algorithm has a low computational cost when compared with EKF-SLAM.

Mobile autonomous systems are very important for marine scientific investigation and military applications. Many algorithms have been studied to deal with the computational efficiency problem required for large scale Simultaneous Localization and Mapping (SLAM) and its related accuracy and consistency. Among these methods, submap-based SLAM is a more effective one. By combining the strength of two popular mapping algorithms, the Rao-Blackwellised particle filter (RBPF) and extended information filter (EIF), this paper presents a Combined SLAM—an efficient submap-based solution to the SLAM problem in a large scale environment. RBPF-SLAM is used to produce local maps, which are periodically fused into an EIF-SLAM algorithm. RBPF-SLAM can avoid linearization of the robot model during operating and provide a robust data association, while EIF-SLAM can improve the whole computational speed, and avoid the tendency of RBPF-SLAM to be over-confident. In order to further improve the computational speed in a real time environment, a binary-tree-based decision-making strategy is introduced. Simulation experiments show that the proposed Combined SLAM algorithm significantly outperforms currently existing algorithms in terms of accuracy and consistency, as well as the computing efficiency. Finally, the Combined SLAM algorithm is experimentally validated in a real environment by using the Victoria Park dataset.

In this paper, a feedback control mechanism and its optimization for rotating disk vibration/flutter via changes of air-coupled pressure generated using piezoelectric patch actuators are studied. A thin disk rotates in an enclosure, which is equipped with a feedback control loop consisting of a micro-sensor, a signal processor, a power amplifier, and several piezoelectric (PZT) actuator patches distributed on the cover of the enclosure. The actuator patches are mounted on the inner or the outer surfaces of the enclosure to produce necessary control force required through the airflow around the disk. The control mechanism for rotating disk flutter using enclosure surfaces bonded with sensors and piezoelectric actuators is thoroughly studied through analytical simulations. The sensor output is used to determine the amount of input to the actuator for controlling the response of the disk in a closed loop configuration. The dynamic stability of the disk-enclosure system, together with the feedback control loop, is analyzed as a complex eigenvalue problem, which is solved using Galerkin’s discretization procedure. The results show that the disk flutter can be reduced effectively with proper configurations of the control gain and the phase shift through the actuations of PZT patches. The effectiveness of different feedback control methods in altering system characteristics and system response has been investigated. The control capability, in terms of control gain, phase shift, and especially the physical configuration of actuator patches, are also evaluated by calculating the complex eigenvalues and the maximum displacement produced by the actuators. To achieve a optimal control performance, sizes, positions and shapes of PZT patches used need to be optimized and such optimization has been achieved through numerical simulations.

The five-link parallel mechanism is proposed to improve joint bionic performance, and the kinematics is established for the closed chain joint actuated by two antagonistic artificial pneumatic muscles (PMs). Interference and singularity constraints are analyzed, and the joint torque model is given based on the spring-damp dynamics. Through extracting the spring force term from torque equations, the compliance of bionic joint is derived and expressed as the ratio of angle to spring torque. Energy consumption is analyzed using the PM length varying. Based on MATLAB/SimMechanics, the relationships between the axil installation parameters and swing performances are illustrated through the simulations, including the effect of the installation height and width varying on the angle scope, swing response, compliance, and energy consumption. The bionic shoulder and elbow joints are optimally designed. Compared to the conventional joints, the swing angular range of the proposed joints is enhanced, and the contraction amount of PMs is reduced. The optimal mechanism is more humanoid.

Landslide susceptibility mapping (LSM) predicts potential landslide risks based on historical landslides, playing a crucial role in landslide prevention and control. This study proposes a novel approach for LSM in open-pit mines, leveraging a comprehensive landslide case database. This method addresses the limitations of traditional LSM approaches, such as sensitivity to historical landslide records, low-quality inventories, and limited dataset size. By expanding the training set and incorporating case-based reasoning (CB-LSM) using Long Short-Term Memory (LSTM) for machine learning, the model demonstrated a 10% improvement in predictive accuracy compared to statistical models, showing strong applicability even in areas with low-quality landslide inventories. Additionally, this study enhances the practical utility of LSM by introducing a similar case matching method, employing Fuzzy C-Means (FCM) clustering, which links high-risk areas identified by LSM with specific landslide prevention measures. The proposed approach not only improves the accuracy of landslide predictions but also strengthens the connection between LSM results and effective landslide prevention strategies. Given the higher risks associated with landslides in open-pit mines, the proposed method significantly lowers the threshold for applying LSM in these environments, offering a valuable tool for safeguarding personnel and property.

The Fushun West Open-pit Mine, recognized as Asia’s largest open-pit coal mine, transitioned to a closure management phase following the cessation of operations in 2019. This study presents a quantitative analysis of the water inflow sources by meticulously reviewing historical data on the mine’s drainage system and the hydrogeological conditions of the mining area. A numerical model representing the unsaturated-saturated groundwater dynamics was constructed to illustrate the spatial distribution of the groundwater flow. Employing the water storage analysis module within the WebGIS system, projections were made regarding the natural water accumulation process at varying water levels. The results indicate that, under natural conditions, an estimated 48 years would be required to reach a water level elevation of 50 m. However, this timeframe could be drastically reduced to approximately 3.3 years by implementing an accelerated filling strategy that channels water from the Hun River. Numerical simulations elucidate the failure mechanism of the water-sensitive slope on the northern bank during the water storage process, as well as the potential impact on the slopes above the water level. This comprehensive approach provides critical insights for managing decommissioned open-pit mines and optimizing water storage schemes.

Surface subsidence is common geologic hazard in coal mines when longwall mining method is applied. In this research, a Bonded Block Model (BBM) is selected to numerically simulate and analyze the surface subsidence of coal mines. Parametric study is conducted and key factors controlling the subsidence are characterized. BBM models with no bedding planes realistically simulate the maximum subsidence, but overestimate the subsidence levels above the panel edges. The bedding planes introduced in the BBM models potentially provide more realistic subsidence profile. The separation of the planes and downward bending of the layered strata, as commonly observed in coal mines, are simulated and captured. In addition, high horizontal stress plays a key role in simulating more realistic subsidence profiles. The methodologies and conclusions of this research are expected to improve our understandings of the surface subsidence of coal mines and provide reliable tools to analyze and characterize this challenging issue.